Exercise Moved Two Measures, Not Three

What the research found

This is a systematic review and meta-analysis, published in BMC Public Health in 2025 by Ruinan Liu and colleagues. A meta-analysis does not run a new experiment. It gathers the results of trials that already exist and combines them, so that many small studies can be read as one larger picture. Here the authors pooled 16 randomised controlled trials involving 685 school-aged children with a diagnosis of ADHD, aged 6 to 13. In each trial, a group of children given an exercise programme was compared with a control group that was not.

The pooled results point in a consistent direction for two of the three abilities examined. For inhibitory control, the capacity to pause and stop an automatic response rather than acting on impulse, exercise produced a moderate improvement (Hedges’ g = 0.60, 95% CI 0.34 to 0.87, p < 0.001). For working memory, the ability to hold information in mind and use it, the effect was also moderate (g = 0.51, 95% CI 0.30 to 0.70, p < 0.001). An effect size around 0.5 to 0.6 is a genuine, visible difference between groups, though not a dramatic one. The confidence intervals stay well clear of zero, which means the finding is unlikely to be a fluke of the numbers.

The third ability, cognitive flexibility, the capacity to switch between tasks or rules, is where the picture becomes murky. The result depended entirely on how the outcome was scored. Under one scoring approach it appeared significant; under another it did not (g = 0.28, p = 0.10, which fails the usual threshold). When a finding flips depending on a bookkeeping choice, the honest reading is that the evidence for cognitive flexibility is inconclusive. The authors treat it that way, and so should we.

Two further details matter. The working memory finding was quite consistent across studies (heterogeneity I-squared = 24.1%), meaning the trials broadly agreed. The inhibitory control finding was less consistent (I-squared = 63.6%), meaning the trials varied more in what they found, and the pooled figure smooths over that spread. And every one of the 16 trials carried some risk of bias, chiefly because the randomisation methods were unclear and, understandably, neither children nor assessors could be blinded to whether a child had been exercising.

What it means for you and your child

The plain summary is encouraging but bounded. Across these trials, structured exercise was associated with real gains in two core executive skills that ADHD often disrupts: the ability to resist impulses and the ability to hold information in mind. Because these were randomised trials comparing exercise against a control condition, the design does support a causal reading. Exercise appears to do something, not merely to accompany children who were doing well anyway.

But be careful about how far this stretches. A meta-analysis inherits the weaknesses of the studies it contains, and here those studies were small and imperfectly controlled. The authors are candid that small samples and methodological limits temper their conclusions. They did not apply a formal grading of the overall quality of evidence, which means the certainty behind these numbers is moderate at best, not settled fact. A pooled effect size is an average across many different children and programmes. It does not promise the same result for your child.

The study also cannot tell you what kind of exercise to choose. The trials mostly used short programmes, under eight weeks, at low to moderate intensity, but the authors did not systematically compare exercise types, intensities, or durations against one another. So the research supports the general idea that regular physical activity helps, without specifying a dose. It says movement matters; it does not say which sport, how hard, or for how long.

There is a further honest caveat. The review included only studies published in Chinese and English, which can skew the picture, and the improvements were measured on laboratory-style attention and memory tasks. A better score on a Stroop or Go/No-Go task is a reasonable proxy for executive function, but it is not the same as calmer mornings, finished homework, or fewer conflicts at the dinner table. The bridge from a test score to daily life is plausible, not guaranteed.

What can you take from it? Regular exercise is low-risk, broadly beneficial for children, and now has reasonable evidence behind it as a support for attention and memory in ADHD. It is a sensible addition to a child’s routine. It is not, on this evidence, a replacement for whatever else is working, whether that is behavioural support, school accommodations, or medication decided with a clinician. Treat exercise as a genuine ally with a modest, well-documented effect, and keep your expectations proportionate to what a handful of small trials can actually show.